I almost forgot a couple of examples. I feel compelled to discuss this because I've written so many "awful" scripts to parse variable html. I have encountered few cases where a human readable answer is the only thing you want ( what will you do with most results? Feed them to another program usually.)
The example I always mention is eutils, http://eutils.ncbi.nlm.nih.gov/entrez/query/static/eutils_help.html The above link explains that this differs from the web interface and shows a simple API. While investigating a similar problem for SEC filings, I was pointed to more complicated protocols such as SOAP, or various things based on exchange of XML data ( not something I think is suitable for simple tasks, but it is an example with a lot of canned support ). To give you an idea of the parsing problem, I also happened to have made available a script to reverse the above mentioned SEC html here, http://www.spottext.com/marchywka/distroform.cfm?src=bb If you go to the doc page, you will get some idea of what it does and you can download the source code to appreciate the problem ( I couldn't even get bash or PERL to work in reasonable time, I ended up writing C++ code .... ). If you don't have a stable machine interface and just want to write a single GUI, you should go through a check list to make sure this is what you really want. Where does the input come from ( typing is subject to error, even cut/paste can miss a leading or trailing character, etc), do I only produce a single result that can be easily translated into a decision or a manscript with no further effort, etc. For that matter, how will you exhaustively test your interface? Even a GUI should have a way to inject commands from an automated process and have a result another process can verify. Maybe it would help if I knew specifically what you are doing but even visual data (see my molecule viewer at the above link) can be scanned prgammatically. For example, I have an auto resize function to take the non-black area and zoom an image to fill the available space to produce a series of images without user intervention ( you could write a script to scan a bunch of pdb files and write them in some given way without producing a gui, the user can then look at the series of bmp files using existing photo viewers and go back and use the GUI for the few interesting cases ). I almost always expect that the graphical output will be for debugging. You can usually pipe text data into data display programs that know how to display publication quality graphics. This isn't always the case but if it is not, you probably won't find a collection of javascript widgets that are all that useful either. And, in any case, the browser is well accessed with java using code that can be moved with no effort into a stand alone application. Java already supports graphics and there are even existing molecule viewing applets etc that may do much of what you need. But, you can't even really think about portability until you have defined somekind of ( machine-to-machine) interface. >> Date: Sun, 13 Apr 2008 17:23:30 -0700 >> From: [EMAIL PROTECTED] >> To: [email protected] >> Subject: Re: [BiO BB] Advanced web query interfaces for biological databases >> >> Thanks for your suggestion, Mike. As a developer I also value the >> concept of an open and freely available API, and I wish that more >> sites would provide these, but I am not sure it is yet appropriate for >> our database as it will be some time before we have a sizable pool of >> users, letalone developers. >> >> Also, I might not be clear on what you mean by "programmatically," but >> I think it is a bit short-sighted to limit the scope of web-based >> interfaces to only casual queries. I am not usually a supporter of the >> term "Web 2.0" because I think it has lately signified little more >> than colorful buttons, but I do believe more and more tasks can be >> accomplished from within the browser. I believe energy should be put >> into moving more advanced tasks online because it removes learning >> curves and hardware restrictions, thereby increasing your base of >> users. >> >> I think it is not easy to develop an intuitive web-based UI that is >> not fairly restrictive, but that does not mean it should not be >> pursued. >> >> Keith >> > _________________________________________________________________ > Going green? See the top 12 foods to eat organic. > http://green.msn.com/galleries/photos/photos.aspx?gid=164&ocid=T003MSN51N1653A > _______________________________________________ > BBB mailing list > [email protected] > http://www.bioinformatics.org/mailman/listinfo/bbb _________________________________________________________________ Get in touch in an instant. Get Windows Live Messenger now. http://www.windowslive.com/messenger/overview.html?ocid=TXT_TAGLM_WL_Refresh_getintouch_042008 _______________________________________________ BBB mailing list [email protected] http://www.bioinformatics.org/mailman/listinfo/bbb
